基于PLS估计的我国商业银行效率影响因素研究

在商业银行效率影响因素研究中引入基于高维投影思想的非参数方法——偏最小二乘方法,并在建模过程中弥补了已有研究未控制时间趋势和宏观经济变量而导致模型不显著和变量符号反常的不足.结果表明:针对本文少样本、自变量存在严重共线性的情况,偏最小二乘法提取的3个成分解释力达到0.984 173,并具有良好的预测性能.财务指标等内生变量对不同产权银行的效率影响是有明显差异的,而产权结构、宏观经济等外生变量的影响是无差异的.

Study on the Influencing Factors of Chinese Commercial Banks Efficiency Based on PLS Method

This paper used non-parametric Partial Least Square method to study the influencing factors of commercial banks efficiency. Compare with the previous researches, this paper controlled time trend and exogenous variables in the model, so the model is more significant and the variable symbols are normal. The empirical results show that, even for the case that the samples were fewer and there were a serious multicollinearity as presented in this article, the explanatory power of the three components extracted by the partial least squares method achieved 0.984 173 and had good predictive performance. Financial variables make a different impact on the efficiency of state-owned and joint-stock commercial banks, but exogenous factors such as structure of property rights and macroeconomic make no difference.